KGTLIR:An Air Target Intention Recognition Model Based on Knowledge Graph and Deep Learning

As a core part of battlefield situational awareness,air target intention recognition plays an important role in modern air operations.Aiming at the problems of insufficient feature extraction and misclassification in intention recognition,this paper designs an air target intention recognition method(KGTLIR)based on Knowledge Graph and Deep Learning.Firstly,the intention recognition model based on Deep Learning is constructed to mine the temporal relationship of intention features using dilated causal convolution and the spatial relationship of intention features using a graph attention mechanism.Meanwhile,the accuracy,recall,and F1-score after iteration are introduced to dynamically adjust the sample weights to reduce the probability of misclassification.After that,an intention recognition model based on Knowledge Graph is constructed to predict the probability of the occurrence of different intentions of the target.Finally,the results of the two models are fused by evidence theory to obtain the target’s operational intention.Experiments show that the intention recognition accuracy of the KGTLIRmodel can reach 98.48%,which is not only better than most of the air target intention recognition methods,but also demonstrates better interpretability and trustworthiness.

Integrated causal inference modeling uncovers novel causal factors and potential therapeutic targets of Qingjin Yiqi granules for chronic fatigue syndrome

Objective:Chronic fatigue syndrome(CFS)is a prevalent symptom of post-coronavirus disease 2019(COVID-19)and is associated with unclear disease mechanisms.The herbal medicine Qingjin Yiqi granules(QJYQ)constitute a clinically approved formula for treating post-COVID-19;however,its potential as a drug target for treating CFS remains largely unknown.This study aimed to identify novel causal factors for CFS and elucidate the potential targets and pharmacological mechanisms of action of QJYQ in treating CFS.Methods:This prospective cohort analysis included 4,212 adults aged≥65 years who were followed up for 7 years with 435 incident CFS cases.Causal modeling and multivariate logistic regression analysis were performed to identify the potential causal determinants of CFS.A proteome-wide,two-sample Mendelian randomization(MR)analysis was employed to explore the proteins associated with the identified causal factors of CFS,which may serve as potential drug targets.Furthermore,we performed a virtual screening analysis to assess the binding affinity between the bioactive compounds in QJYQ and CFS-associated proteins.Results:Among 4,212 participants(47.5%men)with a median age of 69 years(interquartile range:69–70 years)enrolled in 2004,435 developed CFS by 2011.Causal graph analysis with multivariate logistic regression identified frequent cough(odds ratio:1.74,95%confidence interval[CI]:1.15–2.63)and insomnia(odds ratio:2.59,95%CI:1.77–3.79)as novel causal factors of CFS.Proteome-wide MR analysis revealed that the upregulation of endothelial cell-selective adhesion molecule(ESAM)was causally linked to both chronic cough(odds ratio:1.019,95%CI:1.012–1.026,P=2.75 e^(−05))and insomnia(odds ratio:1.015,95%CI:1.008–1.022,P=4.40 e^(−08))in CFS.The major bioactive compounds of QJYQ,ginsenoside Rb2(docking score:−6.03)and RG4(docking score:−6.15),bound to ESAM with high affinity based on virtual screening.Conclusions:Our integrated analytical framework combining epidemiological,genetic,and in silico data provides a novel strategy for elucidating complex disease mechanisms,such as CFS,and informing models of action of traditional Chinese medicines,such as QJYQ.Further validation in animal models is warranted to confirm the potential pharmacological effects of QJYQ on ESAM and as a treatment for CFS.

基于改进causality graph的分布式可伸缩事件关联机制

传统事件关联技术无法有效满足分布式网络管理的需求,本文提出一种分布式可伸缩事件关联机制,采用先分布再集中的关联模式与自适应可伸缩的关联关系。定义了本地关联和网络关联两个过程,首先由设备进行本地关联,而后各地关联结果汇总到管理平台进行网络关联;将事件的关联关系与管理任务的关联关系相结合,根据管理任务在设备端的动态配置情况构建自适应可伸缩的关联关系,并支持对逻辑事件的推理。同时,在改进Causality Graph算法的基础上提出了实现该机制的相关算法。原型系统的应用实例验证了机制的有效性和优越性。

Optimization of a dynamic uncertain causality graph for fault diagnosis in nuclear power plant

Fault diagnostics is important for safe operation of nuclear power plants(NPPs). In recent years, data-driven approaches have been proposed and implemented to tackle the problem, e.g., neural networks, fuzzy and neurofuzzy approaches, support vector machine, K-nearest neighbor classifiers and inference methodologies. Among these methods, dynamic uncertain causality graph(DUCG)has been proved effective in many practical cases. However, the causal graph construction behind the DUCG is complicate and, in many cases, results redundant on the symptoms needed to correctly classify the fault. In this paper, we propose a method to simplify causal graph construction in an automatic way. The method consists in transforming the expert knowledge-based DCUG into a fuzzy decision tree(FDT) by extracting from the DUCG a fuzzy rule base that resumes the used symptoms at the basis of the FDT. Genetic algorithm(GA) is, then, used for the optimization of the FDT, by performing a wrapper search around the FDT: the set of symptoms selected during the iterative search are taken as the best set of symptoms for the diagnosis of the faults that can occur in the system. The effectiveness of the approach is shown with respect to a DUCG model initially built to diagnose 23 faults originally using 262 symptoms of Unit-1 in the Ningde NPP of the China Guangdong Nuclear Power Corporation. The results show that the FDT, with GA-optimized symptoms and diagnosis strategy, can drive the construction of DUCG and lower the computational burden without loss of accuracy in diagnosis.

Local-to-Global Causal Reasoning for Cross-Document Relation Extraction

Cross-document relation extraction(RE),as an extension of information extraction,requires integrating information from multiple documents retrieved from open domains with a large number of irrelevant or confusing noisy texts.Previous studies focus on the attention mechanism to construct the connection between different text features through semantic similarity.However,similarity-based methods cannot distinguish valid information from highly similar retrieved documents well.How to design an effective algorithm to implement aggregated reasoning in confusing information with similar features still remains an open issue.To address this problem,we design a novel local-toglobal causal reasoning(LGCR)network for cross-document RE,which enables efficient distinguishing,filtering and global reasoning on complex information from a causal perspective.Specifically,we propose a local causal estimation algorithm to estimate the causal effect,which is the first trial to use the causal reasoning independent of feature similarity to distinguish between confusing and valid information in cross-document RE.Furthermore,based on the causal effect,we propose a causality guided global reasoning algorithm to filter the confusing information and achieve global reasoning.Experimental results under the closed and the open settings of the large-scale dataset Cod RED demonstrate our LGCR network significantly outperforms the state-ofthe-art methods and validate the effectiveness of causal reasoning in confusing information processing.

基于时序图卷积的动态网络链路预测

动态网络链路预测广泛的应用前景,使得其逐渐成为网络科学研究的热点.动态网络链路演化过程中具有复杂的空间相关性和时间依赖性,导致其链路预测任务极具挑战.提出一个基于时序图卷积的动态网络链路预测模型(dynamic network link prediction based on sequential graph convolution, DNLP-SGC).针对网络快照序列不能有效反映动态网络连续性的问题,采用边缘触发机制对原始网络权重矩阵进行修正,弥补了离散快照表示动态网络存在时序信息丢失的不足.从网络演化过程出发,综合考虑节点间的特征相似性以及历史交互信息,采用时序图卷积提取动态网络中节点的特征,该方法融合了节点时空依赖关系.进一步,采用因果卷积网络捕获网络演化过程中潜在的全局时序特征,实现动态网络链路预测.在2个真实的网络数据集上的实验结果表明,DNLP-SGC在precision, recall, AUC指标上均优于对比的基线模型.

融合异质层次因果图的尾矿堆积坝安全状态分析

为探究尾矿堆积坝内外部因素对堆积坝体安全状态的影响,提出基于异质层次图的尾矿堆积坝安全状态分析方法。该方法通过构建异质层次因果图,将环境、渗流场和应力场等关键因素与堆积坝安全状态建立联系,并结合异质节点属性特征,形成环-渗-应安全状态分析指标体系;通过提出定量指标动态等级区间计算方法和安全状态等级计算模型,将坝体安全状态定性问题的模糊性转化为定量化表达,实现对尾矿堆积坝安全等级评价;以洛阳某尾矿坝为例验证模型的科学性。结果表明:该模型能合理定量分析因素与状态之间联系,并找出堆积坝负向演化的敏感性因素,给后续筑坝过程的安全管理提供决策依据。

动态不确定因果图在中医诊断中的应用探讨

动态不确定因果图(DUCG)已成为中医药领域新兴的、先进的知识表示与推理模型。为更好地应用DUCG为中医临床提供诊断推理与决策支持,在归纳总结DUCG中医药领域研究与应用情况的基础上,分析现阶段DUCG在中医诊断中存在的主要问题,包括中医术语规范统一和中医药知识库质量问题、推理算法和模型建造的方法选择与设计问题、DUCG中医诊断模型的平台化和产品化问题等,并据此展开应用思路与方法探讨,提出应深挖DUCG的技术内涵,根据临床实际需求选择精准、高效的推理建模方法,建立符合中医药理论思想、具有中医特色的智能辅助诊断模型,加强DUCG协同研究平台及产品的开发应用。

基于提示增强与双图注意力网络的复杂因果关系抽取

针对复杂因果句实体密度高、句式冗长等特点导致的外部信息不足和信息传递遗忘问题,提出一种基于提示增强与双图注意力网络(BiGAT)的复杂因果关系抽取模型PE-BiGAT(PromptEnhancementandBi-Graph Attention Network)。首先,抽取句子中的结果实体并与提示学习模板组成提示信息,再通过外部知识库增强提示信息;其次,将提示信息输入BiGAT,同时结合关注层与句法和语义依存图,并利用双仿射注意力机制缓解特征重叠的情况,增强模型对关系特征的感知能力;最后,用分类器迭代预测句子中的所有因果实体,并通过评分函数分析句子中所有的因果对。在SemEval-2010 task 8和AltLex数据集上的实验结果表明,与RPA-GCN(Relationship Position and Attention-Graph Convolutional Network)相比,所提模型的F1值提高了1.65个百分点,其中在链式因果和多因果句中分别提高了2.16和4.77个百分点,验证了所提模型在处理复杂因果句时更具优势。

因果图表征的网络攻击数据集构建

高级可持续威胁攻击因其多阶段可持续的特性,已经成为现阶段网络攻击的主要形式。针对此类型攻击的检测、预测研究,不可避免地需要相关数据集的支撑。在构建数据集时,往往需要真实的网络与主机数据。但出于隐私与安全的考虑,很少有满足要求的开源数据集。现有的数据集也往往只提供原始的网络流和日志数据,对长时攻击上下文解析的缺乏导致单纯地利用神经网络进行数据包的恶性甄别和预测的实用性不足。为了解决这些问题,该文基于网络环境的真实攻击过程数据,构建并公布了一个基于因果图的网络攻击数据集。与传统的原始网络流和日志数据集相比,该数据集充分挖掘了攻击上下文中的因果关系,可以跨长时域对高级可持续威胁攻击进行建模,方便研究人员进行攻击检测与预测的研究。该数据集开源在https://github.com/GuangmingZhu/CausalGraphAPTDataset上。

基于因果反馈的缺失数据集因果关系发现

因果关系发现是因果推断的重要部分,其目标是揭示数据内在的生成机制,并以有向无环图的形式表示.然而关于因果关系发现,现有方法很少考虑到观察数据存在缺失值的情况.在实际场景中,大量数据集存在缺失值,因此估计缺失数据集中的因果关系成为一个亟需解决的问题.本文提出了一种新的基于因果反馈的算法实现关于缺失数据集的因果关系发现,其中生成对抗网络被用于估计缺失数据集的分布,并利用基于Actor-Critic的因果关系发现模块搜索最优因果图,设计了一个基于扩展贝叶斯信息准则的自定义奖励函数,引入分类误差引导模型加速探索过程,提升模型稳定性.在模拟数据和真实数据上进行的大量实验结果表明,本文提出的方法在不同数据缺失率下优于现有方法 .

网络服务异常事件告警因果图构造方法

网络服务系统中,异常事件的发生经常导致系统中产生大量告警事件,形成告警风暴。运维人员需要花费大量的时间和精力从这些告警数据中寻找关键信息、确定异常事件的根源。为了减少运维人员所需处理的告警数量,智能化、自动化地提取告警风暴中的根源告警,基于网络服务告警的传播模式分析,提出了一种告警因果图构造方法,并将其应用于提取异常事件发生时的告警风暴关键信息。实验使用运营商现网管理系统的真实数据集,通过告警风暴摘要提取实验,验证了告警因果图生成的效果,并进行了相关案例的物理意义分析。结果表明,使用告警因果图生成的方式进行告警风暴摘要提取,达到了96%的召回率,保留了绝大部分关键信息。同时,使用该方法对系统产生的告警进行压缩,对较难压缩的告警码的压缩率能够达到66.5%。

A Kind of Fuzzy Causal Diagnosis Method

Aim To improve the causal diagnosis method presented by Bandekar and propose a new method of finding the root fault order according to the fault possibility by means of numerical calculation. Methods Based on the causal graph, by utilization of fuzzified threshold value and fuzzy discrimination matrix, a kind of fuzzy causal diagnosis method was given and the fault possibility of each elements in the root fault candidate set (RFCS) was obtained. Results and Conclusion The order of each element in the RFCS can be obtained by the fault possibility, which makes the location of fault much easier. The diagnosis speed of this method is quite high, and by means of the fuzzified threshold value and fuzzy discrimination matrix, the result is more robust to noises and bad parameter's choice.

有向无环图的构建规范:ESC-DAGs方法介绍及脑血管病研究案例解读

有向无环图(directed acyclic graph,DAG)是构建因果推断中因果假设的可视化工具,可通过后门准则筛选需要校正的最小校正子集,控制观察性研究中的混杂偏倚。然而现有研究在绘制DAG时缺乏规范,文献中使用的DAG质量良莠不齐,在应用上存在困境。ESC-DAGs方法的出现为构建DAG提供了标准化流程。该方法基于系统综述的基本思想广泛检索相关文献,并将构建过程分为映射(mapping)、转化(translation)和整合(integration)三个阶段:首先基于检索出来的每项研究构建饱和DAG,然后根据希尔准则和潜在结果框架筛选图中的有向边,最后将所有保留的有向边整合在一张DAG中。ESC-DAGs方法充分利用了文献信息和研究者的专业知识,为DAG的应用提供了合理可行的标准,具有重要的指导意义。本文将介绍ESC-DAGs方法的基本内容,并通过卒中临床研究示例展示构建DAG的具体过程。

基于动态不确定因果图的航天器故障诊断方法

针对航天器智能化故障诊断的问题,基于动态不确定因果图(Dynamic Uncertain Causality Graph,DUCG)构建诊断模型,克服了基于规则的方法、数据驱动方法存在的诊断正确率低、数据依赖程度高、可解释性差等问题。DUCG基于领域专家的经验知识、以图形化的方式表达航天器遥测参数与可能的故障之间的不确定性知识,不依赖于已有的故障数据,具有诊断正确率高、可解释性强等特征。使用DUCG构建包含42个故障、129个遥测参数的诊断模型,试验结果表明模型的准确率为100%。

动态不确定因果图(DUCG)在鞍区疾病中的诊断价值

目的拟在采取综合考虑鞍区疾病相关特征的基础上,提出一种基于动态不确定因果图(DUCG)的诊断建模与推理方法。方法通过专科医生和人工智能工程师合作,基于DUCG的算法及原理,融合专家知识、临床经验、医学研究成果、临床样本统计数据多种数据和知识来源,构建基于DUCG的鞍区疾病诊断模型。从北京协和医院电子信息病例系统中选择17种最常见的鞍区疾病相关病例,验证诊断模型的的诊断能力,并以图形方式直观展示疾病和症状之间因果关系的推理过程。结果本组DUCG鞍区疾病诊断模型包含138个变量。采用1个病例进行诊断模型测试,结果表明:颅咽管瘤的概率排名第一,占89.004%,并可得到DUCG推导图,清楚解释疾病病因和所有相关因素,症状和监测结果,及其关联性。本组共测试139例鞍区病例,诊断模型的平均准确率为94%。结论基于DUCG的鞍区疾病诊断模型应用于医学诊断中可弥补个人知识和专业限制,降低误诊率,减轻公共卫生保健资源的负担。

基于强化因果发现的航班地面保障事件影响因素分析

对航班地面保障事件执行日志展开分析以及对国内外航班延误报告进行研究,针对航班地面保障事件执行效率低下的影响因素发现问题,明确关联与因果的统一定义,给出量化影响因素的时间差参考量,提出一种面向实际保障过程的基于增剪边规则的强化因果发现算法,通过最大化奖励分数得到最优的影响因素与目标事件因果图。实验结果表明,提出方法能有效得到真实因果图,在FDR、TPR、SHD与生成的因果图的合理性上超过其它对比方法。

基于DUCG的核电站故障诊断专家系统的设计与实现

专家系统有着广泛的应用领域和应用方式。为增强核电站系统的安全性,协助操作员在故障发生时进行分析决策,建立了基于动态不确定因果图(DUCG)理论的核电站故障诊断专家系统。文章详细介绍了这一系统,重点介绍了系统知识库、推理机和图形化用户界面的设计与功能实现过程。验证结果表明,这一系统能进行准确的故障诊断,显示故障的发展过程,为操作员决策提供支持。

去除推荐场景多混淆因子的因果去偏方法

如何去除推荐场景中存在的偏差问题,是推荐系统提升效果的重大挑战。现有的推荐模型只是拟合数据,基于相关性去除偏差会受到虚假相关性的影响;基于因果关系去除偏差则由于场景复杂很难抽取更加全面的因果关系。因此,尽可能多地考虑各因素之间的因果关系并去除偏差问题很有必要。该文从因果角度对用户行为在物品分类的分布不平衡和物品类别流行度两个混淆因子在推荐流程中的因果关系进行研究,提出考虑多因素的去混淆方法,有效去除推荐中的偏差问题。首先,分析各变量之间的因果关系并构建因果图;其次,使用前门调整和后门调整去除两个混淆因子造成的虚假相关性以及偏差;最后,将该方法应用在神经网络因子分解机上,在两个公开数据集上进行了实验并验证。从仿真实验结果可知,该方法相比于目前的最优方法都有不同程度的提升。

知识与句法融合的因果关系抽取网络

因果关系抽取作为关系抽取的一个重要任务,近年来得到了广泛关注。现有的因果关系抽取方法大多将句法结构和背景知识割裂开进行研究,早期的因果关系抽取方法偏重于从句法结构层面进行分析,随着深度学习技术的发展,预训练模型结合背景知识的方法成为主流。然而上述两种方法均未完全融合句内信息和外部知识,带来了不同程度的信息缺失。为了解决这一问题,提出了结合句法结构和背景知识的因果关系抽取模型。该模型将句子解析为同时包含句法和知识的知识句法图结构,使用图卷积网络进行信息融合。模型同时考虑了句法和知识两部分信息,从而进一步丰富了实体嵌入,达到了良好的因果关系抽取效果。本模型在EventStoryLine数据集上取得了良好效果,F1值达到0.445,与现有方法相比提高了2.3%。